CN103339938B - Perform the motion vector prediction of video coding - Google Patents

Abstract

In general, this disclosure describes techniques for performing motion vector prediction for video coding. An apparatus including a motion compensation unit can implement the techniques. The motion compensation unit determines a spatial candidate motion vector MVP associated with a current portion of a video frame and prunes the spatial candidate motion vector to remove duplicates but does not remove temporal candidate motion vectors. the motion compensation unit selects one of the temporal candidate motion vector or the spatial candidate motion vector remaining after pruning as the selected candidate motion vector based on a motion vector predictor MVP index signaled in the bitstream, And performing motion compensation based on the selected candidate motion vector.

Description

Perform motion vector prediction for video decoding

This application asserts U.S. Provisional Application Nos. 61/436,997, filed January 27, 2011, U.S. Provisional Application Nos. 61/449,985, filed March 7, 2011, and 61 /561,601, each of which is hereby incorporated by reference in its entirety.

technical field

This disclosure relates to video coding, and more particularly to motion compensation aspects of video coding.

Background technique

Digital video capabilities can be incorporated into a wide variety of devices, including digital television, digital direct broadcast systems, wireless broadcast systems, personal digital assistants (PDAs), laptop or desktop computers, digital cameras, digital recording devices , digital media players, video game devices, video game consoles, cellular or satellite radio phones, video teleconferencing devices, etc. Digital video devices implement video compression techniques such as the standards defined by MPEG-2, MPEG-4, ITU-TH.263, ITU-T H.264/MPEG-4 Part 10 Advanced Video Coding (AVC) and the like An extension of the standard to describe video compression techniques to transmit and receive digital video information more efficiently. New video coding standards are being developed, such as the High Efficiency Video Coding (HEVC) standard being developed by the Joint Collaborative Team-Video Coding (JCT-VC), a collaboration between MPEG and ITU-T . The emerging HEVC standard is sometimes referred to as H.265, but this name is not yet official.

Contents of the invention

In general, this disclosure describes techniques for specifying motion vector predictors (MVPs). MVP is often used in video coding as a way to increase the efficiency with which motion compensation is performed. Rather than performing a search for a block matching the current block in a reference frame, the video encoder may select a motion vector for the current block from the MVP list. In some examples, the MVP list may include motion vectors for four blocks that are spatially adjacent to the current block, as well as motion vectors for co-located blocks from reference frames that temporally precede or succeed the current frame. A selected one of the MVPs is then used for the current block, reducing, if not eliminating, the process of motion compensation.

In one example, a method of encoding video data includes determining a spatial candidate motion vector associated with a current portion of a current video frame, wherein the spatial candidate motion vector includes a motion vector for the current video frame associated with the current portion motion vectors determined from adjacent neighboring portions; pruning the spatial candidate motion vectors to remove at least one of the spatial candidate motion vectors; and determining a time associated with the current portion of the current video frame Candidate motion vectors. The temporal candidate motion vectors include motion vectors determined for portions of reference video frames. The method further comprises: selecting one of the temporal candidate motion vector or the spatial candidate motion vector remaining after pruning as a selected candidate motion vector; and signaling the selected candidate in a bitstream motion vector.

In another example, an apparatus for encoding video data includes means for determining a spatial candidate motion vector associated with a current portion of a current video frame, wherein the spatial candidate motion vector includes A motion vector determined from a neighboring portion of a frame adjacent to the current portion; means for pruning the spatial candidate motion vectors to remove at least one of the spatial candidate motion vectors; means for determining temporal candidate motion vectors associated with said current portion of said current video frame. The temporal candidate motion vectors include motion vectors determined for portions of reference video frames. The apparatus further comprises: means for selecting one of the temporal candidate motion vector or the spatial candidate motion vector remaining after pruning as the selected candidate motion vector; and for signaling in a bitstream means for the selected candidate motion vector.

In another example, an apparatus for encoding video data includes a motion compensation unit that determines a spatial candidate motion vector associated with a current portion of a current video frame, wherein the spatial candidate motion vector includes motion vectors determined from adjacent portions of a frame adjacent to the current portion; pruning the spatial candidate motion vectors to remove at least one of the spatial candidate motion vectors; and determining all The temporal candidate motion vector associated with the current part. The temporal candidate motion vectors include motion vectors determined for portions of reference video frames. The apparatus also includes: a mode selection unit that selects one of the temporal candidate motion vector or the spatial candidate motion vector remaining after pruning as a selected candidate motion vector; and an entropy coding unit that, in bits The selected candidate motion vector is signaled in a stream.

In another example, a non-transitory computer-readable medium includes instructions that, when executed, cause one or more processors to: determine a spatial candidate motion associated with a current portion of a current video frame vector, wherein the spatial candidate motion vector comprises a motion vector determined for a neighboring portion of the current video frame adjacent to the current portion; pruning the spatial candidate motion vector to remove one of the spatial candidate motion vectors at least one of: determining a temporal candidate motion vector associated with the current portion of the current video frame, wherein the temporal candidate motion vector comprises a motion vector determined for a portion of a reference video frame; selecting the temporal candidate the motion vector or one of the spatial candidate motion vectors remaining after pruning as a selected candidate motion vector; and signaling the selected candidate motion vector in a bitstream.

In another example, a method of decoding video data includes determining a spatial candidate motion vector associated with a current portion of a current video frame, wherein the spatial candidate motion vector includes a motion vector for the current video frame corresponding to the current and pruning the spatial candidate motion vectors to remove at least one of the spatial candidate motion vectors but not for the current portion of the current video frame Determine the temporal candidate motion vectors. The temporal candidate motion vector comprises a motion vector determined for a portion of the reference video frame that is at the same position in the reference video frame as the current portion is in the current video frame. The method also includes selecting one of the temporal candidate motion vectors or one of the spatial candidate motion vectors remaining after pruning based on a motion vector predictor (MVP) index signaled in the bitstream being a selected candidate motion vector; and performing motion compensation based on the selected candidate motion vector.

In another example, an apparatus for decoding video data includes means for determining a spatial candidate motion vector associated with a current portion of a current video frame, wherein the spatial candidate motion vector includes motion vectors determined for adjacent portions of a frame adjacent to the current portion; and for pruning the spatial candidate motion vectors to remove at least one of the spatial candidate motion vectors but not for the current means for determining temporal candidate motion vectors for said current portion of a video frame. The temporal candidate motion vectors include motion vectors determined for portions of reference video frames. The apparatus also includes means for selecting one of the temporal candidate motion vectors or one of the spatial candidate motion vectors remaining after pruning based on a motion vector predictor (MVP) index signaled in a bitstream. one as means for selecting a candidate motion vector; and means for performing motion compensation based on the selected candidate motion vector.

In another example, an apparatus for decoding video data includes a motion compensation unit that determines a spatial candidate motion vector associated with a current portion of a current video frame, wherein the spatial candidate motion vector includes motion vectors determined from adjacent portions of a frame adjacent to the current portion; pruning the spatial candidate motion vectors to remove at least one of the spatial candidate motion vectors but not removing the motion vector for the current video frame a temporal candidate motion vector determined for the current portion, wherein the temporal candidate motion vector comprises a motion vector determined for a portion of a reference video frame; based on a motion vector predictor (MVP) index signaled in the bitstream selecting one of the temporal candidate motion vectors or one of the spatial candidate motion vectors remaining after pruning as a selected candidate motion vector; and performing motion compensation based on the selected candidate motion vector.

In another example, a non-transitory computer-readable medium includes instructions that, when executed, cause one or more processors to: determine a spatial candidate motion associated with a current portion of a current video frame vector, wherein the spatial candidate motion vector comprises a motion vector determined for a neighboring portion of the current video frame adjacent to the current portion; pruning the spatial candidate motion vector to remove one of the spatial candidate motion vectors At least one of but not removing temporal candidate motion vectors determined for the current portion of the current video frame, wherein the temporal candidate motion vectors include motion vectors determined for portions of a reference video frame; based on in-bit a motion vector predictor (MVP) index signaled in the stream to select one of the temporal candidate motion vectors or one of the spatial candidate motion vectors remaining after pruning as the selected candidate motion vector; and Motion compensation is performed based on the selected candidate motion vectors.

The details of one or more examples are set forth in the accompanying drawings and the description below. Other features, objects, and advantages will be apparent from the description and drawings, and from the claims.

Description of drawings

1 is a block diagram illustrating an example video encoding and decoding system that may be configured to utilize the techniques described in this disclosure for specifying motion vector predictors (MVPs).

2 is a block diagram illustrating an example of a video encoder that may implement the techniques described in this disclosure for specifying motion vector predictors.

3 is a block diagram illustrating an example of a video decoder that implements the motion vector prediction techniques described in this disclosure.

4 is a flow diagram illustrating exemplary operation of a video encoder in performing the motion vector prediction techniques described in this disclosure.

5 is a flow diagram illustrating exemplary operation of a video decoder in implementing the motion vector prediction techniques described in this disclosure.

6 is a diagram illustrating an exemplary arrangement of adjacent neighboring PUs and temporally co-located PUs of a current prediction unit (PU).

detailed description

Embodiments of the techniques described in this disclosure enable video encoders to specify MVPs in a robust and efficient manner by pruning redundant spatial MVPs, but not including temporally co-located MVPs during the pruning process. In other words, the technique forms an MVP intermediate list containing only spatial MVPs, performs pruning relative to this MVP intermediate list, and then adds temporally co-located MVPs to the pruned MVP intermediate list to form the pruned MVP list. In this way, the loss of a reference frame for a given temporally co-located MVP may not prevent the parsing of the bitstream, which is common in conventional systems, and the coding efficiency gain achieved by applying the pruning process may still be maintained .

1 is a block diagram illustrating an example video encoding and decoding system 10 that may be configured to utilize the techniques described in this disclosure for specifying motion vector predictors (MVPs). As shown in the example of FIG. 1 , system 10 includes source device 12 that generates encoded video for decoding by destination device 14 . Source device 12 may transmit the encoded video to destination device 14 via communication channel 16 or may store the encoded video on storage medium 34 or file server 36 so that the encoded video may be accessed by destination device 14 as needed. . Source device 12 and destination device 14 may comprise any of a wide variety of devices, including desktop computers, notebook (i.e., laptop) computers, tablet computers, set-top boxes, telephone handsets (including cellular phones or handhelds and so-called smartphones), televisions, cameras, display devices, digital media players, video game consoles, etc.

In many cases, such devices may be equipped for wireless communication. Accordingly, communication channel 16 may comprise a wireless channel. Alternatively, communication channel 16 may comprise a wired channel, a combination of wireless and wired channels, or any other type of communication channel or communication channel suitable for transmitting encoded video data (e.g., a radio frequency (RF) spectrum or one or more combination of physical transmission lines). In some examples, communication channel 16 may form part of a packet-based network such as a local area network (LAN), a wide area network (WAN), or a global network such as the Internet. Accordingly, communication channel 16 generally represents any suitable communication medium, or collection of different communication media, for transmitting video data from source device 12 to destination device 14, including any suitable combination of wired or wireless media. Communication channel 16 may include routers, switches, base stations, or any other equipment that may be used to facilitate communication from source device 12 to destination device 14 .

As further shown in the example of FIG. 1 , source device 12 includes video source 18 , video encoder 20 , modulator/demodulator 22 (“modem 22 ”), and transmitter 24 . In source device 12, video source 18 may include a source such as a video capture device. For example, a video capture device may include one or more of a video camera, a video archive containing previously captured video, a video feed interface for receiving video from a video content provider, and/or for generating Computer graphics data with computer graphics systems as source video. As one example, if video source 18 is a video camera, source device 12 and destination device 14 may form a so-called camera phone or video phone. However, the techniques described in this disclosure are not limited to wireless applications or settings, and are applicable to non-wireless devices that include video encoding and/or decoding capabilities. Accordingly, source device 12 and destination device 14 are merely examples of coding devices that may support the techniques described herein.

Video encoder 20 may encode captured, pre-captured, or computer-generated video. Once encoded, video encoder 20 may output this encoded video to modem 22 . Modem 22 may then modulate the encoded video according to a communication standard, such as a wireless communication protocol, after which transmitter 24 may transmit the modulated encoded video data to destination device 14 . Modem 22 may include various mixers, filters, amplifiers, or other components designed for signal modulation. Transmitter 24 may include circuits designed for transmitting data, including amplifiers, filters, and one or more antennas.

Captured, pre-captured, or computer-generated video encoded by video encoder 20 may also be stored onto storage media 34 or file server 36 for later retrieval, decoding, and use. Storage medium 34 may include a Blu-ray Disc, DVD, CD-ROM, flash memory, or any other suitable digital storage medium for storing encoded video. Destination device 14 may access encoded video stored on storage medium 34 or file server 36, decode the encoded video to generate decoded video, and replay the decoded video.

File server 36 may be any type of server capable of storing and transmitting encoded video to destination device 14 . Example file servers include a web server (e.g., for a website), an FTP server, a network attached storage (NAS) device, a local disk drive, or any other type capable of storing and transmitting encoded video data to a destination device. device. The transmission of encoded video data from file server 36 may be a streaming transmission, a download transmission, or a combination of both. Destination device 14 may access file server 36 via any standard data connection, including an Internet connection. This connection can include a wireless channel (for example, a Wi-Fi connection or a wireless cellular data connection), a wired connection (for example, DSL, cable modem, etc.), a combination of both wired and wireless channels, or be suitable for accessing storage Any other type of communication channel for encoded video data on a file server.

In the example of FIG. 1 , destination device 14 includes receiver 26 , modem 28 , video decoder 30 , and display device 32 . Receiver 26 of destination device 14 receives the information via channel 16 and modem 28 demodulates the information to produce a demodulated bitstream for video decoder 30 . Information communicated via channel 16 may include a variety of syntax information generated by video encoder 20 for use by video decoder 30 in decoding the associated encoded video data. This syntax may also be included within encoded video data stored on storage media 34 or file server 36 . Each of video encoder 20 and video decoder 30 may form part of a respective encoder-decoder (CODEC) capable of encoding or decoding video data.

Display device 32 of destination device 14 represents any type of display capable of presenting video data for viewing by a viewer. Although shown as being integrated with destination device 14 , display device 32 may be integrated with destination device 14 or external to destination device 14 . In some examples, destination device 14 may include an integrated display device and may also be configured to interface with an external display device. In other examples, destination device 14 may be a display device. In general, display device 32 displays decoded video data to a user and may include any of a variety of display devices, such as a liquid crystal display (LCD), a plasma display, an organic light emitting diode (OLED) display, or another type of display device. display device.

This disclosure may generally involve video encoder 20 "signaling" syntax information to another device (eg, video decoder 30 ). It should be understood, however, that video encoder 20 may signal information by associating syntax elements with various encoded portions of video data. That is, video encoder 20 may "signal" data by storing certain syntax elements to headers of various encoded portions of video data. In some cases, such syntax elements may be encoded and stored (eg, to storage medium 34 or file server 36 ) prior to being received and decoded by video decoder 30 . Thus, the term "signaling" may generally refer to the transfer of syntax or other data used to decode compressed video data, whether such communication occurs in real-time or near real-time or over a span of time, such as may occur at the time of encoding Occurs when syntax elements are stored to a medium, which can then be retrieved by a decoding device at any time after storage to this medium.

Video encoder 20 and video decoder 30 may operate according to a video compression standard, such as the High Efficiency Video Coding (HEVC) standard currently under development, and may conform to the HEVC test model (HM). Alternatively, video encoder 20 and video decoder 30 may be based on other proprietary or industry standards such as the ITU-T H.264 standard, otherwise known as MPEG-4 Part 10 Advanced Video Coding (AVC) or such standards The expansion operates. However, the techniques of this disclosure are not limited to any particular coding standard. Other examples include MPEG-2 and ITU-T H.263.

HM refers to a block of video data as a coding unit (CU). In general, a CU has a similar purpose to a macroblock coded according to H.264, except that the CU does not have the size distinction associated with a macroblock of H.264. Accordingly, a CU may be split into sub-CUs. In general, a reference to a CU in this disclosure may refer to a largest coding unit (LCU) of a picture or a sub-CU of an LCU. For example, syntax data within a bitstream may define an LCU, which is the largest coding unit in terms of number of pixels. A LCU can be split into sub-CUs, and each sub-CU can be split into sub-CUs. Syntax data for the bitstream may define the maximum number of times an LCU can be split, referred to as the maximum CU depth. Accordingly, a bitstream may also define a smallest coding unit (SCU).

LCUs may be associated with a hierarchical quadtree data structure. In general, a quadtree data structure includes one node for each CU, where the root node corresponds to an LCU. If a CU is split into four sub-CUs, the node corresponding to the CU contains references for each of the four nodes corresponding to the sub-CU. Each node of the quadtree data structure may provide syntax data for the corresponding CU. For example, a node in a quadtree may include a split flag indicating whether the CU corresponding to the node is split into sub-CUs. Syntax elements for a CU may be defined recursively, and may depend on whether the CU is split into sub-CUs.

An unsplit CU may include one or more prediction units (PUs). In general, a PU represents all or a portion of a corresponding CU and includes data used to retrieve reference samples for the PU. For example, when a PU is encoded in intra-mode, the PU may include data describing the intra-prediction mode for the PU. As another example, when a PU is encoded in inter mode, the PU may include data defining one or more motion vectors for the PU. A motion vector typically identifies a co-located CU in one or more reference frames, where the term "reference frame" refers to a frame that temporally occurs before or after the frame in which the PU resides. Data defining a PU with respect to a CU may also describe, for example, partitioning of the CU into one or more PUs. The partition mode may differ between whether the CU is uncoded, intra-prediction mode encoded, or inter-prediction mode encoded.

Data defining a motion vector may describe, for example, the horizontal component of the motion vector, the vertical component of the motion vector, the resolution of the motion vector (e.g., quarter-pixel precision or one-eighth pixel precision), the reference to which the motion vector points frame, a prediction direction identifying whether the identified reference frame is before or after the current frame, and/or a reference list (eg, list 0 or list 1 ) for motion vectors. Alternatively, the data defining a motion vector may describe the motion vector in terms of something called a motion vector predictor (MVP). A motion vector predictor may include motion vectors for neighboring PUs or temporally co-located PUs. Typically, a list of five MVPs is formed in a defined manner (e.g., listed starting with the MVP with the largest amplitude to the MVP with the smallest amplitude (i.e., the largest or smallest displacement between the current PU to be coded and the reference PU) MVP, or list MVPs based on position (i.e., top block, left block, corner block, temporal block), where four of the five MVPs are spatial MVPs selected from four adjacent PUs and the fifth MVP is the selected The temporally co-located MVP from the temporally co-located PU in the reference frame.

Although typically the temporal candidate motion vectors are at the same location in the same portion of the reference frame relative to the motion vectors of the current portion in the current frame, the techniques should not be strictly limited to temporal candidate motion vectors at the same location. Rather, the techniques may be implemented with respect to any temporal candidate motion vector, whether or not the temporal candidate motion vectors are co-located. In some cases, the video encoder may identify a temporal candidate motion vector that is not co-located with the current block or portion of the current frame and select this temporal candidate motion vector as the temporal MVP. Typically, the video encoder may signal that a temporal MVP that is not co-located is used, or in some cases a given context may indicate that a temporal MVP that is not co-located is used (in which case the video encoder may not signal whether to select time MVP who is not in the same position).

After forming the list of five MVPs, video encoder 20 may evaluate each of the MVPs to determine which MVP provides the best match for a given rate-to-distortion profile selected for encoding video. Best rate and distortion characteristics. Video encoder 20 may perform a rate-distortion optimization (RDO) procedure with respect to each of the five MVPs, selecting the one of the MVPs with the best RDO result. Alternatively, video encoder 20 may select one of the five MVPs stored to the list that best approximates the motion vector determined for the current PU.

In any case, video encoder 20 may specify the motion vector using data comprising: an index identifying a selected one of the MVPs in a list of five MVPs, one or more references to which the motion vector points Frames (often in list form), and a prediction direction identifying whether the prediction is unidirectional or bidirectional. Alternatively, the data defining the motion vector may only specify the index of the selected MVP in the list of five MVPs without specifying the reference frame and prediction direction, which indicates to the video decoder that the selected one of the MVPs will be used exclusively for the current PU .

In addition to having one or more PUs that define one or more motion vectors, a CU may also include one or more transform units (TUs). After prediction using a PU, a video encoder may calculate residual values for the portion of the CU corresponding to the PU, where such residual values may also be referred to as residual data. The residual values may be transformed, quantized and scanned. A TU is not necessarily limited to the size of a PU. Accordingly, a TU may be larger or smaller than the corresponding PU of the same CU. In some examples, the maximum size of a TU may be the size of a corresponding CU. This disclosure also uses the term "block" to refer to any one or combination of CUs, PUs, and/or TUs.

In general, encoded video data may include predictive data and residual data. Video encoder 20 may generate prediction data during intra-prediction mode or inter-prediction mode. Intra prediction generally involves predicting pixel values in a block of the same picture relative to reference samples in neighboring previously coded blocks of the same picture. Inter prediction generally involves predicting pixel values in a block of a picture relative to data of previously coded pictures.

Following intra-prediction or inter-prediction, video encoder 20 may calculate residual pixel values for the block. The residual value generally corresponds to the difference between the block's predicted pixel value data and the block's real pixel value data. For example, residual values may include pixel difference values indicating differences between coded pixels and predicted pixels. In some examples, coded pixels may be associated with a pixel block to be coded, and predicted pixels may be associated with one or more pixel blocks used to predict the coded block.

To further compress the residual values of a block, the residual values may be transformed into a set of transform coefficients that compress as much data (also referred to as "energy") as possible into as few coefficients as possible. Transformation techniques may include discrete cosine transform (DCT) processes or conceptually similar processes, integer transforms, wavelet transforms, or other types of transforms. A transform converts the residual value of a pixel from the spatial domain to the transform domain. The transform coefficients correspond to a two-dimensional matrix of coefficients, usually of the same size as the original block. In other words, there are exactly as many transform coefficients as there are pixels in the original block. However, due to the transform, many of the transform coefficients may have values equal to zero.

Video encoder 20 may then quantize the transform coefficients to further compress the video data. Quantization generally involves mapping values in a relatively large range to values in a relatively small range, thus reducing the amount of data required to represent quantized transform coefficients. More specifically, quantization may be applied according to a quantization parameter (QP), which may be defined at the LCU level. Thus, the same level of quantization may apply to all transform coefficients in TUs associated with different PUs of a CU within an LCU. However, rather than signaling the QP itself, a change in QP (ie, Δ) may be signaled with the LCU. ΔQP defines the change of the LCU's quantization parameter relative to some reference QP (eg, the previously transmitted QP of the LCU).

After quantization, video encoder 20 may scan the transform coefficients, generating a one-dimensional vector from a two-dimensional matrix including the quantized transform coefficients. Video encoder 20 may then perform statistical lossless encoding (which is often misreferred to as "entropy encoding") to encode the resulting array to compress the data even further. In general, entropy coding includes one or more processes that collectively compress a sequence of quantized transform coefficients and/or other syntax information. For example, syntax elements such as ΔQP, prediction vectors, coding modes, filters, offsets, or other information may also be included in the entropy coded bitstream. The scanned coefficients along with any syntax information are then entropy coded, eg, via content adaptive variable length coding (CAVLC), context adaptive binary arithmetic coding (CABAC), or any other statistical lossless coding process.

As noted above, the data defining motion vectors for a PU may take numerous forms. Video encoder 20 may implement different ways in which motion vectors may be represented to compress motion vector data. Video encoder 20 may implement a mode referred to as merge mode to represent motion vectors as indices identifying MVPs stored to an MVP list constructed in a defined manner. In implementing the reverse process of this merge mode, video decoder 30 receives this index, reconstructs the list of five MVPs according to a defined manner, and selects the one of the five MVPs in the list indicated by the index. Video decoder 30 then instantiates the selected one of the MVPs as a motion vector for the associated PU with the same resolution as the selected one of the MVPs and pointing to the same direction as the selected one of the MVPs. The same reference frame as the reference frame. When implementing merge mode, video encoder 20 may not need to perform motion estimation to the maximum extent necessary to derive motion vectors, specifying the horizontal and vertical components of motion vectors, motion vector resolution, motion vector direction (meaning, A motion vector points to a frame temporally preceding or following the current frame) or a reference frame index, thereby potentially reducing the processor cycles required to determine the motion vector and compressing the motion vector data.

Video encoder 20 may also implement an adaptive motion vector prediction (AMVP) mode, which, similar to merge mode, also includes representing motion vectors as indexes identifying MVPs stored to an MVP list constructed in a defined manner. However, in contrast to merge mode, video encoder 20 may also specify prediction directions and reference frames, effectively overriding these portions of a selected one of the MVPs. When implementing AMVP mode, video encoder 20 may not need to perform motion estimation to the maximum extent necessary to obtain motion vectors, need not specify the horizontal and vertical components of motion vectors and motion vector resolution, thereby potentially reducing the need for determining motion vectors. The required processor loops and compresses the motion vector data.

With the development of various coding standards, even more efficient representations of motion vectors have been developed. For example, proposals for the emerging HEVC standard have suggested ways in which MVP indexes can be compressed through a process called "pruning" or "MVP pruning." In performing this pruning process, video encoder 20 builds a list of five MVPs in a defined manner and then prunes or removes any redundant MVPs. That is, video encoder 20 may remove any MVPs that have the same amplitude on both the X and Y components and refer to the same reference frame, where these MVPs are considered "redundant MVPs" in this disclosure. Alternatively, video encoder 20 may only add "unique" MVPs to the list, meaning that these MVPs have a different amplitude and/or reference to a different reference in the X and Y directions than all other MVPs already contained in the list. frame. Whether pruned after being added to the list or when the list is created, the pruning process can reduce the size of the list, with the result that fewer bits can be used to signal or otherwise specify a selected one of the MVPs, which is because shorter lists generally require a smaller number of bits to represent the largest index value.

For example, assume for illustration purposes that none of the five MVPs are pruned. In this case, the video encoder may signal the index into this list of five MVPs using a truncated unary code comprising up to four bits to indicate the particular MVP to be selected. In the proposal, the truncated unary code used to signal the selection of the fifth MVP in the list of five MVPs is 1111, the truncated unary code used to signal the selection of the fourth MVP in the list of five MVPs is 1111. The short unary code is 1110 to signal the selection of the third MVP in the list of five MVPs The truncated unary code is 110 to signal the selection of the second MVP in the list of five MVPs The truncated unary code for is 10, and the truncated unary code used to signal the selection of the first MVP in the list of five MVPs is 0. However, if the MVP list can be pruned to three MVPs (meaning that two of the MVPs are redundant), then video encoder 20 can use a truncated unary code that consumes up to two bits (e.g., where code 11 can be used Signaling the third MVP), thereby potentially saving compared to cases where no pruning is used or pruning is not possible (e.g. when there is no redundant MVP) and the fifth or fourth MVP in the list of five MVPs is selected one. Thus, to some extent, the code depends on the size of the MVP list, where a smaller MVP list (meaning a list with fewer MVPs) results in a smaller code (meaning the code requires fewer bits to represent list of selected MVPs).

While pruning may improve coding efficiency by reducing the code length used to signal the index of the selected MVP in the MVP list, such pruning may also affect the ability of video decoder 30 to successfully parse the bitstream. Because the code depends on the size of the pruned MVP list, video decoder 30 needs to know the number of MVPs in the pruned MVP list. However, when the reference frame where the temporally co-located MVP is located is lost, the temporally co-located MVP is unavailable and the video decoder 30 cannot determine whether the temporally co-located MVP is unique or redundant. Therefore, video decoder 30 cannot determine whether the pruned MVP list already includes this temporally co-located MVP, and thus cannot determine the size of the pruned MVP list. Without being able to determine the size of the pruned MVP list, video decoder 30 is then unable to determine the maximum length of the code, which in turn prevents video decoder 30 from parsing the code from the bitstream.

According to the techniques described in this disclosure, video encoder 20 may specify MVPs in a potentially robust and efficient manner by pruning redundant spatial MVPs, but not including temporally co-located MVPs during the pruning process. In other words, video encoder 20 may implement the techniques described in this disclosure to form an intermediate list of MVPs containing only spatial MVPs, perform pruning relative to this intermediate list of MVPs, and then combine temporally co-located MVPs with the pruned MVPs. The MVP intermediate list to form a pruned MVP list. In this way, the loss of a reference frame for a given temporally co-located MVP may not prevent dissection of the bitstream, which is common in conventional systems, while still maintaining the coding efficiency gain achieved by using the pruning process at least some of the .

To illustrate, video encoder 20 first determines a spatial candidate motion vector associated with a current portion (eg, a CU) of a current video frame. Furthermore, the spatial candidate motion vectors include neighboring motion vectors determined for neighboring PUs associated with the corresponding CU that are neighboring to the current PU. Typically, these neighboring PUs are positioned adjacent to the current PU to the left, top left, directly above, and above right, as shown in more detail with respect to the example of FIG. 6 . Video encoder 20 uses these spatial candidate motion vectors because they have been determined for the blocks. Given that video encoder 20 typically performs motion estimation/compensation from top-to-bottom and left-to-right, for any blocks positioned directly to the right or directly below the current PU, video encoder 20 still needs to compute motion vectors for those blocks . However, while described with respect to these spatial motion vectors, the techniques may be implemented in video encoder 20 that performs motion estimation/compensation in a different order (eg, top-to-bottom, right-to-left). Additionally, the techniques may be implemented with respect to more or fewer spatial or temporal motion vectors.

After determining these spatial motion vectors, video encoder 20 then prunes the spatial candidate motion vectors to remove duplicates in the spatial candidate motion vectors. Video encoder 20 may identify repeated spatial candidate motion vectors as any of the candidate spatial motion vectors that have the same amplitude for both the x-axis and y-axis components of the candidate motion vector and are from the same reference frame. Video encoder 20 adds a candidate spatial motion vector to be added to this list by removing duplicates from a list, which may be referred to as an intermediate list of spatial candidate motion vectors, or only upon determining that the candidate spatial motion vector is not a duplicate. This intermediate list is used to perform pruning.

After pruning the spatial candidate motion vectors in this manner, video encoder 20 may then determine a temporal candidate motion vector for the current PU of the current video frame. Furthermore, the temporal candidate motion vector includes a motion vector determined for a PU of a reference video frame that is co-located with the current PU of the current video frame. Video encoder 20 may then select one of the temporal candidate motion vector or the spatial candidate motion vector remaining after performing the pruning process as the selected candidate motion vector. Video encoder 20 then signals the selected candidate motion vector in the bitstream.

In some cases, video encoder 20 may determine whether each of the determined spatial candidate motion vectors is spatially predicted or temporally predicted. In other words, the determined spatial candidate motion vectors themselves may be predicted temporally from co-located blocks in the reference frame or spatially from blocks adjacent to the block for which each of the spatial candidate motion vectors was determined . In response to this determination, video encoder 20 may further remove one or more of the determined spatial candidate motion vectors from the pruning process. For example, video encoder 20 may remove from the pruning process those of the spatial candidate motion vectors that are determined to be temporally predicted themselves, because if the reference from which the prediction of these temporally predicted spatial candidate motion vectors is based If part of the frame is lost, then these temporally predicted spatial candidate motion vectors may not be available to the decoder. Video encoder 20 may then select one of the temporal candidate motion vector, the temporally predicted spatial candidate motion vector, or one of the spatially predicted spatial candidate motion vectors remaining after pruning, and signal this selection in the bitstream. selected candidate motion vectors.

Alternatively, video encoder 20 may replace these temporally predicted spatial candidate motion vectors with default candidate motion vectors defining default motion information, rather than removing the temporally predicted spatial candidate motion vectors from the pruning process. This default motion vector information may include, for example, a motion vector amplitude, a prediction direction identifying whether the reference frame precedes or follows the current frame in time, and a reference index identifying the reference frame. Video encoder 20 may select the first available spatially predicted spatial candidate motion vector by taking the average of those available spatially predicted spatial candidate motion vectors, or use a statically configured link between video encoder 20 and video decoder 30 This default motion vector information is determined by the default motion vector information in the user (just to name a few).

By eliminating from the pruning process spatial candidate motion vectors that may be lost or not available to the video decoder 30 (e.g., because of transmission errors of the compressed video data or storage errors at the video encoder 20 or video decoder 30) Candidate motion vectors, video encoder 20 may signal selected candidate motion vectors in a manner that enables video decoder 30 to properly parse bits in the event that these temporally predicted spatial candidate motion vectors are lost or become unavailable flow. Likewise, in an alternative, by replacing temporally predicted spatial candidate motion vectors with default candidate motion vectors, video encoder 20 may signal selected candidate motion vectors in a manner that enables video decoder 30 to Properly dissects the bitstream in case predicted spatial candidate motion vectors are lost or become unavailable.

Typically, video encoder 20 signals the selected candidate motion vector using a unary representing an index of the selected candidate motion vector as arranged in a list. Video encoder 20 may arrange temporal candidate motion vectors and spatial candidate motion vectors remaining after performing the pruning process in a set or defined manner (e.g., from highest amplitude to lowest amplitude, lowest amplitude to highest amplitude, temporal motion vector first and then The remaining spatial motion vectors sorted from highest to lowest amplitude or lowest to highest amplitude, etc.), forming a list of candidate motion vectors. Alternatively, video encoder 20 may signal some identifier that indicates the manner in which the motion vectors are arranged in the list. Regardless, video encoder 20 then identifies one of the candidate motion vectors stored to the list, encoding the index of the selected one of the candidate motion vectors stored to the list using the unary code in the manner described above.

Video decoder 30 receives this bitstream, decodes the index and forms an intermediate list of spatial candidate motion vectors (if available). As described above, motion vector prediction is used at video encoder 20 to encode motion vectors for neighboring blocks and select a temporal motion vector where the reference frame defining this temporal motion vector is lost (e.g., due to memory corruption, bus error, or transmission error), one or more of the spatial candidate motion vectors may not be available. Alternatively, when video encoder 20 encodes the motion vector of this neighboring PU using motion vector prediction and selects one of the spatial motion vectors which is itself motion vector predicted from a temporal motion vector in which the reference frame of this temporal motion vector is defined When lost (eg, due to memory corruption, bus errors, or transmission errors), one or more of the spatial candidate motion vectors may not be available. Video decoder 30 can overcome this problem by removing unavailable temporally predicted candidate motion vectors from the pruning process, or replacing these unavailable temporally predicted spatial candidate motion vectors with default candidate motion vectors in the alternative . In this regard, video decoder 30 implements the techniques described above with respect to video encoder 20 in a substantially similar manner in order to properly parse the signaled selected candidate motion vectors from the bitstream.

In any case, even assuming one or more candidate motion vectors are missing, the techniques enable the MVP to be signaled in a manner such that parsing of the bitstream is facilitated. By ensuring that temporal candidate motion vectors are always present in the list after pruning, video encoder 20 ensures that video decoder 30 can determine the number of available motion vectors and thereby parse the index from the bitstream. Likewise, video encoder 20 ensures that video decoder 30 can determine the number of available motion vectors and The index from the bitstream is then parsed. In this way, even if the slice storing the temporal candidate motion vector and/or the temporally predicted spatial candidate motion vector is lost, video decoder 30 may still parse the bitstream, regardless of whether unary codes are used. In particular, video decoder 30 may parse the bitstream knowing that temporal candidate motion vectors and/or temporally predicted spatial candidate motion vectors are always included in the MVP list and are never pruned from the MVP list. In the alternative where the temporally predicted spatial candidate motion vectors are replaced by default candidate motion vectors, video encoder 20 effectively ensures that such temporally predicted spatial candidate motion vectors are not lost because video decoder 30 is configured in such a way as to always These motion vectors can be determined using the same techniques that video encoder 20 performs to determine this default candidate motion vector.

To illustrate the case where the temporal candidate motion vector is lost, consider the case where the amplitude of the spatial candidate motion vector is 1, 1, 1, and 1 and the amplitude of the temporal candidate motion vector is -1. Video decoder 30 may implement the techniques to form a list initially with only spatial candidate motion vectors (which may be referred to as an MVP list), such that the MVP list is 1, 1, 1, and 1. The decoder then prunes this spatial-only MVP list such that this MVP list is defined as one. The decoder then adds the temporal candidate motion vectors to the MVP list such that the MVP list is defined as -1 and 1. The encoder may then signal mvp_idx of 0 or 1 to indicate that one of these motion vectors is selected (or mvp_idx of 0 or 10 if truncated unary codes are not used). As with the unary codes described above, the techniques of this disclosure remove the possibility of having to infer that only one candidate motion vector is available after pruning, since there will always be at least one spatial candidate motion vector and the temporal candidate spatial motion vector.

In this way, the techniques enable video encoders to specify MVPs in a potentially robust and efficient manner by pruning redundant spatial MVPs, but not considering temporally co-located MVPs during the pruning process. In other words, the technique forms an MVP intermediate list containing only spatial MVPs, performs pruning relative to this MVP intermediate list, and then adds temporally co-located MVPs to the pruned MVP intermediate list to form the pruned MVP list. In this way, the loss of reference frames for a given temporally co-located MVP may not prevent parsing of the bitstream, which is common in conventional systems, while still maintaining the coding efficiency gains achieved by using the pruning process.

In some cases, the techniques may be applied in other situations. For example, the fourth version of the HEVC test model (HM4.0) proposes to prune MVPs first, and then add additional MVPs if the total number of MVPs remaining after pruning is less than five. In other words, HM4.0 will prune five MVPs (ie, one temporal and four spatial) to produce a pruned MVP list. If the number of MVPs in this pruned MVP list is less than five, then HM 4.0 adds non-redundant MVPs until the total number of MVPs in the pruned list equals five. These non-redundant MVPs may be selected from other spatial or temporal blocks or may be generated based on the MVPs in the pruned MVP list (e.g., selecting the y-component of one of the MVPs in the pruned MVP list and x-component of another different MVP of ). In this case, the video encoder may implement the techniques described in this disclosure to select additional non-redundant MVPs such that only spatial MVPs are selected and/or used to generate these additional non-redundant MVPs.

By selecting only the spatial MVPs or using the existing spatial MVPs remaining after pruning to generate these additional non-redundant MVPs, the video encoder can ensure that the video decoder can properly determine the selected one of the MVPs. That is, by always having five MVPs, the video encoder ensures that the video decoder can always parse the MVP index from the bitstream, but if the temporal MVP is missing, the video decoder may not be able to accurately construct the MVP list because when the temporal MVP When missing, the MVPs cannot determine the order of the MVPs relative to each other. The techniques described in this disclosure can reduce or potentially eliminate the impact of missing temporal MVPs by not selecting any temporal MVPs or spatial MVPs that are themselves predicted from temporal MVPs as additional non-redundant MVPs.

The techniques for specifying motion vector prediction described with respect to examples of this disclosure may be applied to video coding to support any of a variety of multimedia applications, such as over-the-air television broadcasts, cable television transmissions, satellite television transmissions, (for example) Streaming video transmission via the Internet, encoding of digital video for storage on a data storage medium, decoding of digital video stored on a data storage medium, or other applications. In some examples, system 10 may be configured to support one-way or two-way video transmission for applications such as video streaming, video playback, video broadcasting, and/or video telephony.

Although not shown in FIG. 1 , in some aspects video encoder 20 and video decoder 30 may be integrated with an audio encoder and decoder, respectively, and may include appropriate multiplexer-demultiplexers ( MUX-DEMUX) unit or other hardware and software to handle the encoding of both audio and video in a common data stream or in separate data streams. If applicable, in some examples the MUX-DEMUX unit may conform to the ITU H.223 multiplexer protocol, or other protocols such as User Datagram Protocol (UDP).

Video encoder 20 and video decoder 30 may each be implemented as any of a variety of suitable encoder circuits, such as one or more microprocessors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field Programmable gate array (FPGA), discrete logic, software, hardware, firmware, or any combination thereof. When the techniques are implemented partially in software, a device may store instructions for the software in a suitable non-transitory computer-readable medium and execute the instructions in hardware using one or more processors to Perform the techniques of the present invention. Each of video encoder 20 and video decoder 30 may be included in one or more encoders or decoders, either of which may be integrated as part of a combined encoder/decoder (CODEC) in a respective device middle.

2 is a block diagram illustrating an example of video encoder 20 that may implement techniques for specifying motion vector predictors. Video encoder 20 may perform intra- and inter-coding of blocks within video frames, including macroblocks, or partitions or sub-partitions of macroblocks. Intra-coding relies on spatial prediction to reduce or remove spatial redundancy in video within a given video frame. Inter-coding relies on temporal prediction to reduce or remove temporal redundancy in video within adjacent frames of a video sequence. Intra-mode (I-mode) may refer to any of several spatial-based compression modes, and inter-modes such as unidirectional prediction (P-mode) or bi-directional prediction (B-mode) may refer to several temporal-based compression modes any of the Although components for inter-mode encoding are depicted in FIG. 2, it should be understood that video encoder 20 may further include components for intra-mode encoding. However, for the sake of brevity and clarity, such components have not been illustrated.

As shown in FIG. 2, video encoder 20 receives a current video block within a video frame to be encoded. In the example of FIG. 2 , video encoder 20 includes motion compensation unit 44 , motion estimation unit 42 , memory 64 , summer 50 , transform unit 52 , quantization unit 54 , and entropy coding unit 56 . For video block reconstruction, video encoder 20 also includes inverse quantization unit 58 , inverse transform unit 60 , and summer 62 . A deblocking filter (not shown in FIG. 2 ) may also be included to filter block boundaries to remove blocking artifacts from the reconstructed video. A deblocking filter typically filters the output of summer 62 when required. Although described as including memory 64 generally referring to random access memory (RAM), dynamic RAM (DRAM), static RAM (SRAM), flash memory, or other persistent or non-persistent chip-based storage media, any type of Non-transitory computer readable media, including hard drives, optical drives, magnetic disk drives, etc.

During the encoding process, video encoder 20 receives a video frame or slice to be coded. A frame or slice can be divided into video chunks. Motion estimation unit 42 and motion compensation unit 44 perform inter-predictive coding of the received video block relative to one or more blocks in one or more reference frames to provide temporal compression. Intra-prediction unit 46 may also perform intra-predictive coding of a received video block relative to one or more neighboring blocks in the same frame or slice as the block to be coded to provide spatial compression.

As further shown in the example of FIG. 2 , video encoder 20 also includes mode select unit 40 . Mode select unit 40 may select one of the coding modes (intra or inter), eg, based on error results, and provide the resulting intra-coded or inter-coded block to a summer. 50 to generate residual block data and provided to summer 62 to reconstruct the encoded block for use as a reference frame.

Motion estimation unit 42 and motion compensation unit 44 may be highly integrated, but are illustrated separately for conceptual purposes. Motion estimation is the process of generating motion vectors that estimate the motion of video blocks. A motion vector, for example, may indicate the displacement of a predictive block within a predictive reference frame (or other coded unit) relative to a current block being coded within a current frame (or other coded unit). A predictive block is a block that is found to closely match the block to be coded in terms of pixel difference, which may be determined by sum of absolute difference (SAD), sum of square difference (SSD), or other difference metrics. A motion vector may also indicate the displacement of a partition of a macroblock. Motion compensation may involve retrieving or generating a predictive block based on a motion vector determined by motion estimation. Furthermore, in some examples, motion estimation unit 42 and motion compensation unit 44 may be functionally integrated.

Motion estimation unit 42 calculates a motion vector for the video block of the inter-coded frame by comparing the video block of the inter-coded frame to the video block of the reference frame in memory 64 . Motion compensation unit 44 may also interpolate sub-integer pixels of a reference frame (eg, an I frame or a P frame). The emerging HEVC standard (and the ITU H.264 standard) store reference frames through one or more list data structures (which are often referred to as "lists"). Therefore, the data stored in the memory 64 can also be regarded as a list. Motion estimation unit 42 compares blocks of one or more reference frames (or lists) from memory 64 with blocks to be encoded of a current frame (eg, a P frame or a B frame). When a reference frame in memory 64 includes values for sub-integer pixels, a motion vector calculated by motion estimation unit 42 may refer to a sub-integer pixel location of the reference frame. Motion estimation unit 42 sends the calculated motion vectors to entropy coding unit 56 and motion compensation unit 44 . A reference frame block (which may include a CU) identified by a motion vector may be referred to as a predictive block. Motion compensation unit 44 calculates an error value for a predicted block of a reference frame.

Motion compensation unit 44 may calculate prediction data based on the prediction block. Video encoder 20 forms a residual video block by subtracting the prediction data from motion compensation unit 44 from the original video block being coded. Summer 50 represents the component that performs this subtraction operation. Transform unit 52 applies a transform, such as a discrete cosine transform (DCT) or a conceptually similar transform, to the residual block, producing a video block comprising residual transform coefficient values. Transform unit 52 may perform other transforms that are conceptually similar to DCT, such as the transform defined by the H.264 standard. Wavelet transforms, integer transforms, subband transforms, or other types of transforms may also be used. In any case, transform unit 52 applies a transform to the residual block, producing a block of residual transform coefficients. The transform may convert residual information from a pixel value domain to a transform domain (eg, frequency domain). Quantization unit 54 quantizes the residual transform coefficients to further reduce bit rate. The quantization process may reduce the bit depth associated with some or all of the coefficients. The degree of quantization can be modified by adjusting quantization parameters.

Following quantization, entropy coding unit 56 entropy codes the quantized transform coefficients. For example, entropy coding unit 56 may perform content adaptive variable length coding (CAVLC), context adaptive binary arithmetic coding (CABAC), or another entropy coding technique. Following entropy coding by entropy coding unit 56, the encoded video may be transmitted to another device or archived for later transmission or retrieval. In the case of context-adaptive binary arithmetic coding, the context may be based on neighboring macroblocks.

In some cases, entropy coding unit 56 or another unit of video encoder 20 may be configured to perform other coding functions in addition to entropy coding. For example, entropy coding unit 56 may be configured to determine CBP values for macroblocks and partitions. Also, in some cases, entropy coding unit 56 may perform run-length coding of coefficients in a macroblock or partition thereof. In particular, entropy coding unit 56 may apply a zigzag scan or other scan pattern to scan the transform coefficients in a macroblock or partition, and encode runs of zeros for further compression. Entropy coding unit 56 may also construct header information with appropriate syntax elements for transmission in the encoded video bitstream.

Inverse quantization unit 58 and inverse transform unit 60 apply inverse quantization and inverse transform, respectively, to reconstruct a residual block in the pixel domain, eg, for later use as a reference block. Motion compensation unit 44 may calculate a reference block by adding the residual block to a predictive block of one of the frames of reference frame store in memory 64 . Motion compensation unit 44 may also apply one or more interpolation filters to the reconstructed residual block to calculate sub-integer pixel values for use in motion estimation. Summer 62 adds the reconstructed residual block to the motion compensated prediction block produced by motion compensation unit 44 to produce a reconstructed video block for storage in the reference frame store of memory 64 . The reconstructed video blocks may be used by motion estimation unit 42 and motion compensation unit 44 as reference blocks for inter-coding blocks in subsequent video frames.

As noted above, in some cases motion estimation unit 42 may not compute motion vectors but instead determine a list of motion vector predictors, four of which are spatial candidate motion vectors and the One of the motion vector predictors is a temporal candidate motion vector. In general, motion estimation unit 42 forgoes motion vector calculations in order to reduce the computational complexity of motion estimation and thereby increase the speed at which video data can be encoded while also reducing power consumption. According to the techniques described in this disclosure, motion estimation unit 42 may determine four spatial candidate motion vectors (or fewer after determining that one or more of the spatial candidate motion vectors are themselves temporally predicted, where such temporally predicted the intermediate list of spatial candidate motion vectors to be removed from the pruning process), prune this list of (potentially only spatially predicted) spatial candidate motion vectors, and combine the temporal candidate motion vectors (and potentially temporally predicted spatial candidate motion vector) to the pruned list of (potentially only spatially predicted) spatial candidate motion vectors. Alternatively, motion estimation unit 42 may determine an intermediate list of four (when replacing these temporally predicted spatial candidate motion vectors with default candidate motion vectors) spatial candidate motion vectors, pruning the spatial candidate motion vectors (including ) and add the temporal candidate motion vectors to the pruned list of spatial candidate motion vectors. Motion estimation unit 42 may output this list, including the temporal motion vector candidates and the pruned spatial motion vector candidates, to motion compensation unit 44 .

Motion compensation unit 44 may then identify a reference frame block (again, which may be referred to as a predictive block) for each candidate motion vector included in the list. Motion compensation unit 44 may then calculate predictive data based on the predictive blocks determined for each of the candidate motion vectors. Video encoder 20 may then determine residual data for each prediction data calculated for a corresponding one of the candidate motion vectors, transform the residual data, quantize the transcoded residual data, and then convert the transcoded residual data in the manner described above. The quantized residual data is entropy coded. Video encoder 20 may then perform the inverse operation to decode this entropy-encoded residual data generated relative to the teaching of the remaining candidate motion vectors after pruning, to regenerate the reference data in the form of reconstructed video blocks. Mode select unit 40 may analyze each of the reconstructed video blocks generated relative to each of the candidate motion vectors to select one of the candidate motion vectors. Mode select unit 40 may select one of the candidate motion vectors that provides the best rate-distortion ratio through a process commonly referred to as "rate-distortion optimization," which is often abbreviated as "RDO."

RDO typically involves comparing a reconstructed frame, slice, or block compressed to achieve a certain rate (which typically refers to the bit rate at which compressed video data comprising the compressed frame, slice, or block can be transmitted) with the original frame, slice or block and determine the amount of distortion between the original frame, slice or block and the reconstructed frame, slice or block at a given rate. Mode select unit 40 may encode the same video data using multiple different metrics for achieving or attempting to achieve a given rate, thereby performing the distortion optimization process with respect to these various metrics. In this case, mode select unit 40 may compare the RD output for each reconstructed video block and select the video block that provides the least distortion at the target rate.

Mode select unit 40 may then indicate this selection to motion estimation unit 42, which proceeds to interface with entropy coding unit 56 to inform entropy coding unit 56 of the selection. Typically, motion estimation unit 42 interfaces with entropy coding unit 56 to indicate that motion vector prediction is to be performed along with an index identifying a selected candidate motion vector. As noted above, motion estimation unit 42 may arrange candidate motion vectors in a defined manner, such as by highest amplitude to lowest amplitude, or lowest amplitude to highest amplitude, or in any other defined manner. Alternatively, motion estimation unit 42 may also signal to entropy coding unit 56 the manner in which candidate motion vectors are arranged in this list (which may also be referred to as an MVP list). Entropy coding unit 56 may then use the unary or truncated unary code to encode this index along with any other information that may be necessary to indicate that MVP is performed to encode motion data. Entropy coding unit 56 may output the encoded index as a syntax element in the bitstream (which may be denoted "mvp_idx"), which may be stored or transmitted in the manner described above with respect to the example of FIG. 1 .

In some cases, entropy coding unit 56 performs a type of entropy coding known as context adaptive binary arithmetic coding (CABAC). When performing CABAC, entropy coding unit 56 may select one of a number of so-called contexts (which are different code tables specified for different contexts in order to more efficiently compress the video data associated with the corresponding context) and The compressed residual data is encoded using a context-bound code table. Entropy coding unit 56 may select one of the contexts based on context information, which may include a reference index determined when performing motion vector prediction, a number of unique motion vector candidates, and a prediction determined when performing motion vector prediction. direction.

3 is a block diagram illustrating an example of a video decoder 30 that decodes an encoded video sequence. In the example of FIG. 3 , video decoder 30 includes entropy decoding unit 70 , motion compensation unit 72 , intra prediction unit 74 , inverse quantization unit 76 , inverse transform unit 78 , memory 82 , and summer 80 . In some examples, video decoder 30 may perform a decoding process that is substantially reciprocal to the encoding process described with respect to a video encoder, such as video encoder 20 shown in the examples of FIGS. 1 and 2 . Although generally reciprocal, in some cases video decoder 30 may perform techniques similar to those performed by video encoder 20 . In other words, video decoder 30 may perform a process substantially similar to that performed by video encoder 20 . Furthermore, video encoder 20 may perform video decoding in the process of performing video encoding, as described above. To illustrate, inverse quantization unit 58 , inverse transform unit 60 , and summer 62 of video encoder 20 may perform substantially similar operations as inverse quantization unit 76 , inverse transform unit 78 , and summer 80 of video decoder 30 .

As shown in the example of FIG. 3 , entropy decoding unit 70 receives an encoded bitstream, which, for purposes of illustration, is assumed to include identifying selected candidate motion vectors (where, again, these candidate motion vectors may be referred to as Unary or truncated unary coding index for motion vector predictor (MVP). In performing a process generally reciprocal to entropy coding unit 56 of video encoder 20, entropy decoding unit 70 may receive a syntax element or other coding data for the current PU that indicates that motion vector prediction is performed to determine the The motion vector of the current PU. In response to this syntax element or other coded data, entropy decoding unit 70 determines to implement the techniques described in this disclosure to determine the number of candidate motion vectors remaining after pruning in order to properly parse the unary or truncated unary code from the bitstream.

To determine the number of candidate motion vectors, entropy coding unit 70 may interface with motion compensation unit 72, instructing motion compensation unit 72 to determine the number of candidate motion vectors according to the techniques described in this disclosure. Motion compensation unit 72 retrieves spatial candidate motion vectors for PUs adjacent to the current PU and temporal candidate motion vectors for co-located PUs in the reference frame. Entropy coding unit 70 may provide the reference frame identified for the current PU to motion compensation unit 72 (typically as another syntax element in the bitstream). Alternatively, motion compensation unit 72 may be configured with respect to AMVP or merge mode to identify in a set manner (e.g., backward or forward, eg, one, two, or any other number, from the current frame where the current PU resides) A reference frame is retrieved for temporal candidate motion vectors.

Motion compensation unit 72 may then construct the four spatial candidate motion vectors (or one or the other in determining the spatial candidate motion vectors) in a manner substantially similar to that described above with respect to motion compensation unit 44 of video encoder 20. One or more are themselves temporally predicted (fewer, where such temporally predicted spatial candidate motion vectors are removed from the pruning process) intermediate list, pruning this list of (potentially only spatially predicted) spatial candidate motion vectors, And this pruned list of temporal candidate motion vectors (and potentially temporally predicted spatial candidate motion vectors) is combined with the (potentially only spatially predicted) spatial candidate motion vectors. Alternatively, motion compensation unit 72 may determine the four spatial candidate motion vectors (replacing the temporally predicted spatial candidate with a default candidate motion vector) in a manner, again substantially similar to that described above with respect to motion compensation unit 44 of video encoder 20. motion vectors), pruning this list of spatial candidate motion vectors (including one or more of the determined default candidate motion vectors), and combining the pruned list of temporal and spatial candidate motion vectors. In each case, motion compensation unit 72 outputs this list of candidate motion vectors determined after performing pruning as a pruned MVP list. After generating this pruned MVP list, motion compensation unit 72 counts the number of candidate motion vectors in this list and signals this number to entropy coding unit 70 . Entropy coding unit 70 may then properly parse unary or truncated unary coding indices from the provided bitstream for the reasons described above.

After parsing the unary or truncated unary coding index, entropy coding unit 70 may then decode the unary or truncated unary coding index to generate an index to the MVP list. Entropy coding unit 70 then passes this index to motion compensation unit 72, which then selects the one of the candidate motion vectors identified by the index from the pruned MVP list. For inter-coded blocks, motion compensation unit 72 may then generate inter-prediction data based on the identified motion vectors. Motion compensation unit 72 may use this motion vector to identify a predictive block in a reference frame that is stored to memory 82 . For intra-coded blocks, intra-prediction unit 74 may use the intra-prediction mode received in the bitstream to form a predicted block from spatially adjacent blocks. Inverse quantization unit 76 inverse quantizes (ie, dequantizes) the quantized block coefficients provided in the bitstream and decoded by entropy decoding unit 70 . The inverse quantization process may include, for example, a conventional process as defined by the H.264 decoding standard. The inverse quantization process may also include using the quantization parameter QP _Y calculated by summer 50 for each macroblock to determine the degree of quantization, and likewise determine the degree of inverse quantization that should be applied.

Inverse transform unit 60 applies an inverse transform (eg, inverse DCT, inverse integer transform, or a conceptually similar inverse transform process) to the transform coefficients to produce a residual block in the pixel domain. Motion compensation unit 72 generates motion compensated blocks, possibly performing interpolation based on interpolation filters. Identifiers of interpolation filters to be used for motion estimation with sub-pixel precision may be included in the syntax elements. Motion compensation unit 72 may use interpolation filters as used by video encoder 20 during encoding of the video block to calculate interpolated values for sub-integer pixels of a reference block. Motion compensation unit 72 may determine the interpolation filters used by video encoder 20 from received syntax information and use the interpolation filters to produce predictive blocks.

Motion compensation unit 72 uses some of the syntax information to determine the size of the CU used to encode the frame(s) of the encoded video sequence, partition information describing the manner in which each CU of the frame(s) of the encoded video sequence is partitioned, A mode indicating the manner in which each CU was encoded, one or more reference frames (or lists) for each inter-coded CU, and other information used to decode the encoded video sequence.

Summer 80 sums the residual block with the corresponding prediction block produced by motion compensation unit 72 or intra-prediction unit to form a decoded block. A deblocking filter may also be applied to filter the decoded blocks, if desired, in order to remove blocking artifacts. The decoded video blocks are then stored in memory 82 in a reference frame store (which may be referred to as a decoded picture buffer in the HEVC standard), which provides reference blocks for subsequent motion compensation, and also The decoded video is generated for presentation on a display device (eg, display device 32 of FIG. 1 ).

In some cases, a temporal candidate motion vector may not be available, such as when the slice specifying the temporal candidate motion vector is lost (ie, not recovered or received in the encoded bitstream, as one example). When such a temporal candidate motion vector is not available, motion compensation unit 72 may set this temporal candidate motion vector to a default value or otherwise determine default motion vector information for this temporal candidate motion vector. In some cases, this default motion vector information for temporal candidate motion vectors may be reconstructed depending on whether the reference frame is intra-coded or not. When determining that a reference frame is intra-coded, motion compensation unit 72 may derive a default Default motion vector information for candidate motion vectors. Likewise, one or more of the temporally predicted spatial candidate motion vectors may be unavailable or missing and may be based on reference frames that are at the same position in the reference frame as the current portion is in the current frame Partially determined spatial motion vectors are used to obtain default motion vector information of default candidate motion vectors.

Various aspects of the techniques set forth in this disclosure may also address issues that arise when performing CABAC or any other context-dependent lossless statistical coding process because spatial candidate motion vectors are themselves predicted from missing or lost temporal motion vectors . In accordance with these aspects of the techniques, entropy coding unit 70 may overcome this problem by deactivating missing spatial candidate motion vectors. Alternatively, entropy coding unit 70 may replace the missing spatial candidate motion vector with default motion information. Entropy coding unit 70 may interface with motion compensation unit 72 to determine this default motion information. This default motion information may specify a default motion vector, prediction direction and reference index. In some cases, motion compensation unit 72 is based on the slice type (which indicates whether the current slice is intra-predicted or inter-predicted, etc.), the current CU depth (such as the CU in the above-mentioned quadtree hierarchy where the current PU resides) depth), current PU size, or any other available information to generate this replacement motion information. Motion compensation unit 72 may then provide this default motion information to entropy coding unit 70 . By utilizing this default motion information, entropy coding unit 70 may still perform the CABAC process.

In some instances, the techniques described in this disclosure may also overcome problems when motion compensation unit 72 cannot determine whether a spatial candidate motion vector itself is predicted from a spatial candidate motion vector or a temporal candidate motion vector (such as when its temporal candidate motion vector is missing). ) problems. In these cases where one of the spatial candidate motion vectors is not available, motion compensation unit 72 may implement the techniques of this disclosure to disable spatial motion vector prediction (and then utilize the temporal candidate motion vector at the same location regardless of the motion vector used by the encoder. what the signal is sending). Alternatively, motion compensation unit 72 may determine default motion information in the manner described above.

The technique may further overcome problems that arise when motion compensation unit 72 cannot determine whether a spatial candidate motion vector itself is predicted from a temporal candidate motion vector, such as when its temporal candidate motion vector is lost. In these cases where one of the spatial candidate motion vectors is not available, motion compensation unit 72 may implement the techniques of this disclosure to disable spatial motion vector prediction (and then utilize the temporal candidate motion vector at the same location regardless of the motion vector used by the encoder. what the signal is sending). Alternatively, motion compensation unit 72 may determine default motion information in the manner described above, performing pruning relative to this default motion information for a default candidate motion vector or removing this default candidate motion vector entirely from the pruning process (but specifying it separately to enable bitstream analysis).

As mentioned above, there are two types of motion vector prediction: merge mode and AMVP. For merge mode, motion compensation unit 72 determines motion vector amplitude, prediction direction, and reference index when determining default motion information. For AMVP, motion compensation unit 72 determines the motion vector amplitude, but does not need to determine the prediction direction and reference index, since the prediction direction and reference index are signaled separately in the bitstream for the current PU. Accordingly, motion compensation unit 72 may make the determination of default motion information based on the signaled mode for performing motion vector prediction (ie, for the current PU, whether the signaled type of motion vector prediction is merge mode or AMVP).

4 is a flowchart illustrating exemplary operation of a video encoder, such as video encoder 20 shown in the example of FIG. 2, in performing the motion vector prediction techniques described in this disclosure. Initially, as described above, motion estimation unit 42 may determine a spatial candidate motion vector for the current PU corresponding to the current CU (90). Motion estimation unit 42 may instantiate what may be referred to as an intermediate list or an intermediate spatial MVP list that stores these spatial candidate motion vectors. Motion estimation unit 42 may then prune redundant spatial candidate motion vectors in one of the ways described above (92). In this sense, motion estimation unit 42 may generate an intermediate spatial MVP list of the remaining spatial candidate motion vectors.

After generating this intermediate spatial MVP list of the remaining spatial candidate motion vectors, motion estimation unit 42 may determine the temporal candidate motion vector for the current PU from the co-located PU in the reference frame again as described above (94). Next, motion estimation unit 42 may form an MVP list that includes the remaining spatial and temporal candidate motion vectors (96). Motion estimation unit 42 may provide this MVP list to motion compensation unit 44, which performs motion compensation with respect to each candidate motion vector included in the MVP list in the manner described above. Video encoder 20 then generates residual data based on the prediction data generated by motion compensation performed with respect to each of the candidate motion vectors included in the MVP list. Video encoder 20 applies one or more transforms to the residual data, quantizes the transformed residual data, and then reconstructs the residual data. The reconstructed residual data is then augmented by prediction data generated based on each of the candidate motion vectors included in the MVP list to generate reconstructed video data.

Mode select unit 40 may then select one of the candidate motion vectors from the current PU's MVP list based on the reconstructed video data (98). For example, mode select unit 40 may perform some form of rate-distortion analysis on the reconstructed video data reconstructed relative to each of the candidate motion vectors in the MVP list, and select the candidate motion vectors from the list The one that provides the best rate-distortion metric. Mode select unit 40 may then interface with motion estimation unit 42 to indicate its candidate motion vector selection. Motion estimation unit 42 may then determine an index to the MVP list identifying the selected one of the candidate motion vectors, as described above (100). Motion estimation unit 42 may then provide this index to entropy coding unit 56 . Entropy coding unit 56 may then code an index identifying the selected one of the candidate motion vectors in the MVP list, as described further above (102). Entropy coding unit 56 then inserts the coded index into the bitstream (104).

5 is a flow diagram illustrating exemplary operation of a video decoder, such as video decoder 30 shown in the example of FIG. 3, in implementing the motion vector prediction techniques described in this disclosure. As described above, entropy decoding unit 70 of video decoder 30 initially receives a bitstream that includes a coded index, commonly referred to by its syntax element name as "mvp_idx" or "MVP index" (110). Entropy decoding unit 70 also decodes other syntax elements before or after this MVP index that indicate that the current PU has a motion vector denoted as motion vector prediction. In order to parse this MVP index from the bitstream, entropy decoding unit 70 must first determine the number of candidate motion vectors remaining after performing the pruning process. To determine the number of candidate motion vectors, entropy decoding unit 70 interfaces with motion compensation unit 72, requesting motion compensation unit 72 to provide this number of candidate motion vectors for the current PU.

In response to this request, motion compensation unit 72 determines a spatial candidate motion vector for the current PU in the manner described above (112). If one or more of the spatial candidate motion vectors are not available for reasons set forth in more detail above ("Yes" 114), motion compensation unit 72 may generate motion information in any of the ways described above ( eg default motion information) and perform motion compensation (116, 118) based on the generated motion information. If all spatial candidate motion vectors are available ("No" 114), motion compensation unit 72 prunes redundant spatial candidate motion vectors, as described further above (120).

After pruning redundant spatial candidate motion vectors, motion compensation unit 72 then determines a temporal candidate motion vector for the current PU from the co-located PU in the reference frame as described above (122). If this temporal candidate motion vector is not available for the reasons described above ("YES" 124), motion compensation unit 72 may generate motion information and perform motion compensation based on the generated motion information (116, 118). However, if temporal candidate motion vectors are available ("NO" 124), motion compensation unit 72 forms an MVP list that includes the remaining spatial and temporal candidate motion vectors (126). Motion compensation unit 72 may then determine the number of candidate motion vectors in the MVP list ( 128 ), passing this number to entropy decoding unit 70 .

Entropy decoding unit 70 may then parse the MVP index from the bitstream based on the determined number (130). Entropy decoding unit 70 then decodes the coded MVP index (131). Entropy decoding unit 70 passes the decoded MVP index to motion compensation unit 72, which selects one of the candidate motion vectors from the MVP list based on the decoded MVP index, as described above (132). Motion compensation unit 72 then performs motion compensation based on the selected one of the candidate motion vectors in the manner described above (134). Motion compensation unit 72 may perform motion compensation according to merge mode or AMVP, depending on which mode is signaled in the bitstream or determined by motion compensation unit 72 .

6 is a diagram illustrating an exemplary arrangement of adjacent neighboring PUs 140A-140D of the current PU 144 and the temporally co-located PU 142A. As shown in the example of FIG. 6, current PU 144 is contained within current frame 146A. In time, the current frame 146A is preceded by the reference frame 146B, which is preceded by the reference frame 146C. The adjacent neighbor PU 140A resides spatially adjacent to the left of the current PU 144 . The adjacent neighboring PU 140B resides spatially adjacent to the upper left of the current PU 144 . The adjacent neighboring PU 140C resides spatially adjacently above the current PU 144 . The adjacent neighboring PU 140D resides spatially adjacent to the upper right of the current PU 144 . The temporally co-located PU 142A temporally precedes the current PU 144 and is at the same position within the reference frame 146B as the current PU 144 was located within the current frame 146A.

Each of adjacent neighboring PUs 140A- 140D stores or otherwise provides a spatial candidate motion vector for the current PU 144 , while the temporally co-located PU 142A stores or otherwise provides a temporal candidate motion vector for the current PU 144 . A motion compensation unit of a video decoder (eg, motion compensation unit 72 of video decoder 30 shown in the example of FIG. 2 ) may retrieve these spatial and temporal candidate motion vectors from PUs 140A- 140D and 142A, respectively. Because the temporally co-located PU 142A is contained within a different reference frame 146B than the frame of the current PU 144, this temporally co-located PU 142A is often associated with a different independently decodable portion (often referred to as a slice in the emerging HEVC standard). )Associated. This slice of reference frame 146B may be lost (e.g., in transmission or due to memory or storage corruption) and motion compensation unit 72 may not be able to retrieve this temporally co-located PU 142A that stores the temporal candidate motion vector for the current PU 144 . Losing this temporal candidate motion vector may prevent entropy decoding unit 70 from parsing the bitstream for the reasons described above. The techniques described in this disclosure may enable motion compensation unit 72 to overcome this problem by not including temporal candidate motion vectors in the pruning process.

Likewise, the spatial candidate motion vector for the current PU 144 may be lost when performing MVP to determine the motion vector for one of the neighboring neighboring PUs 140A-140D, with the result that the temporal candidate motion vector is selected and the temporal candidate motion vector is stored The PU at the same location is lost. To illustrate, consider the adjacent neighbor PU 140A, whose temporally co-located PU is identified in the example of FIG. 6 as the temporally co-located PU 142B. If PU 142B is lost and the motion vector of spatially neighboring PU 140A is selected as the temporal candidate motion vector associated with PU 142B, then there is no motion vector information for PU 140A. Therefore, the spatial candidate motion vector for the current PU 144 is also lost. To potentially overcome this missing spatial candidate motion vector, the techniques enable motion compensation unit 72 to generate motion information (eg, default motion information) that may be used as a spatial candidate motion vector for current PU 144 .

Furthermore, such missing spatial candidate motion vectors (or, for that matter, missing temporal candidate motion vectors) may occur when MVP is performed and temporal candidate motion vectors are selected for multiple temporally co-located PUs. To illustrate, assume that MVP is performed to determine a motion vector for PU 142B, PU 142B is temporally co-located with PU 140A, and the temporally co-located PU of PU 142B (ie, PU 142C in the example of FIG. 6 ) is missing. In the absence of the techniques described in this disclosure, this loss would not only potentially prevent the MVP index from being parsed from the bitstream, but would also result in a loss of motion vectors for PU 142B. In the absence of the techniques described in this disclosure, assuming MVP is performed to determine the motion vector for PU 140A and the temporally co-located PU 142B is selected, then the loss of the motion vector for PU 142B results in the loss of the motion vector for PU 140A. The loss of this motion vector affects the current PU 144 because spatial candidate motion vectors are not available. For this reason, the techniques enable motion compensation unit 72 to generate motion information (or, in some cases, to regenerate lost motion information) in order to prevent what may be referred to as the multiple dropout effect.

While the examples above involve removing duplicate spatial candidate motion vectors, the techniques may not necessarily require only duplicate spatial candidate motion vectors to be removed. The techniques may be implemented to perform pruning to substantially remove at least one of the spatial candidate motion vectors. For example, a video encoder may signal at the picture, frame, slice, or block level that the spatial candidate motion vector with the largest or smallest amplitude, to name a few, is to be pruned. Alternatively, the video encoder may signal in the bitstream any criterion by which the MVP may be specified, such as a threshold, as a pruning criterion. In some embodiments, the video encoder and video decoder may agree on a certain profile or other configuration by which candidate motion vectors may be pruned. In some cases, a video decoder may hint when certain candidate motion vectors are to be pruned based on context or other information. Thus, the techniques should not be strictly limited to performing pruning to remove only duplicate spatial candidate motion vectors, but should encompass any technique that can be implemented to prune at least one spatial candidate motion vector.

In one or more examples, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functions may be stored or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media may include computer data storage media or communication media including any medium that facilitates transfer of a computer program from one place to another. Data storage media may be any available media that can be accessed by one or more computers or one or more processors to retrieve instructions, code and/or data structures for implementation of the techniques described in this disclosure. By way of example, and not limitation, such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk or other magnetic storage, flash memory, or may be used to carry or any other medium that stores desired program code in the form of instructions or data structures and that can be accessed by a computer. Also, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then coaxial cable, fiber optic Cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of media. Disk and disc, as used herein, includes compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and blu-ray disc where disks usually reproduce data magnetically, while discs reproduce optically using lasers data. Combinations of the above should also be included within the scope of computer-readable media.

One or more processors such as one or more digital signal processors (DSP), general-purpose microprocessors, application-specific integrated circuits (ASICs), field-programmable logic arrays (FPGAs), or other equivalent integrated or discrete logic circuits to execute the code. Accordingly, the term "processor," as used herein may refer to any of the foregoing structure or any other structure suitable for implementation of the techniques described herein. Additionally, in some aspects, the functionality described herein may be provided within dedicated hardware and/or software modules configured for encoding and decoding or incorporated in a combined codec. Also, the techniques could be fully implemented in one or more circuits or logic elements.

The techniques of this disclosure may be implemented in a wide variety of devices or apparatuses, including wireless handsets, integrated circuits (ICs), or collections of ICs (eg, chipsets). Various components, modules, or units are described in this disclosure to emphasize functional aspects of devices configured to perform the disclosed techniques, but do not necessarily require embodiment by different hardware units. Rather, as described above, the various units may be combined in a codec hardware unit or provided by a collection of interoperable hardware units (comprising one or more processors as described above) in combination with suitable software and/or firmware .

Various examples have been described. These and other examples are within the scope of the following claims.

Claims (61)

1. a method for coding video frequency data, described method comprises:

Determining the spatial candidate motion vector being associated with the current portions of current video frame, wherein said spatial candidate is transported Moving vector includes the motion determining for the adjacent part neighbouring with described current portions of described current video frame Vector；

Determine the time candidate motion vector being associated with the described current portions of described current video frame, when wherein said Between candidate motion vector include the motion vector that determines for the part of reference video frame；

Prune described spatial candidate motion vector with remove repetition person in described spatial candidate motion vector but described Described time candidate motion vector is not considered during pruning process；

One in remaining described spatial candidate motion vector after selecting described time candidate motion vector or pruning As selected candidate motion vector for motion vector prediction process；

Determine the row of remaining described spatial candidate motion vector after identifying described time candidate motion vector and pruning The index of the position of the described selected one in described candidate motion vector in table；And

Send described index with signal in bit stream.

2. method according to claim 1, wherein prunes described spatial candidate motion vector and includes only pruning described space Candidate motion vector is to remove the described repetition person in described spatial candidate motion vector.

3. method according to claim 1,

Wherein said current portions includes current decoding unit CU,

The described part of wherein said reference frame includes the CU of described reference frame.

4. method according to claim 1, it farther includes to produce predicting unit PU comprising prediction data, institute State prediction data and include at least described selected candidate motion vector,

Wherein send described selected candidate motion vector with signal and include in described bit stream described with signal transmission PU。

5. method according to claim 1, wherein sends described index with signal and includes with signal transmission motion vector pre- Measured value MVP indexes, and described MVP index identifies remaining institute after described time candidate motion vector and pruning State the described position of described selected candidate motion vector in the described list of spatial candidate motion vector.

6. method according to claim 1, it farther includes to use unitary code or the one truncating in unitary code to compile Code is described to be indexed to produce encoded index,

Wherein send described index with signal to include sending described encoded index with signal in described bit stream.

7. method according to claim 1, remaining institute after wherein selecting described time candidate motion vector or pruning The one stated in spatial candidate motion vector includes:

Relative to described time candidate motion vector and prune after remaining described spatial candidate motion vector in every One performs rate-distortion analysis；And

Remaining described space after selecting described time candidate motion vector based on described rate-distortion analysis or prune One in candidate motion vector is as described selected candidate motion vector.

8. method according to claim 1, it further comprises determining that in the spatial candidate motion vector of described determination Each is through spatial prediction or through time prediction,

Wherein prune described spatial candidate motion vector to remove the described repetition person in described spatial candidate motion vector It is confirmed as the spatial candidate motion through spatial prediction including only prune in the spatial candidate motion vector of described determination Vector, and do not prune the spatial candidate fortune being confirmed as in the spatial candidate motion vector of described determination through time prediction Any one in moving vector, and

After wherein selecting described time candidate motion vector or pruning in remaining described spatial candidate motion vector One includes selecting described time candidate motion vector, described spatial candidate as described selected candidate motion vector Motion vector is confirmed as through time prediction one or pruning after remaining described spatial candidate motion vector In be confirmed as the one through spatial prediction as described selected candidate motion vector.

9. method according to claim 1, it farther includes:

Determine described determination spatial candidate motion vector in each through spatial prediction still through time prediction；And

Replace with default candidate motion vector and be confirmed as through appointing in the described spatial candidate motion vector of time prediction One, wherein said default candidate motion vector comprises default motions vector information, wherein said default motions vector Information includes motion vector amplitude, identifies described reference frame pre-before or after described present frame in time Survey direction and the reference key identifying described reference frame, and

Wherein prune described spatial candidate motion vector to remove the described repetition person in described spatial candidate motion vector Including prune the one or more of described spatial candidate motion comprising in described default candidate motion vector to Amount is to remove the described repetition person in described spatial candidate motion vector.

10. method according to claim 1, it further comprises determining that it is not through time prediction and to be different from the described time In remaining described spatial candidate motion vector any one one or one after candidate motion vector and pruning Above exceptional space candidate motion vector,

After wherein selecting described time candidate motion vector or pruning in remaining described spatial candidate motion vector One includes selecting described time candidate motion vector, prune after in remaining described spatial candidate motion vector One in one or described exceptional space candidate motion vector is as described selected candidate motion vector.

11. 1 kinds of equipment for coding video frequency data, described equipment includes:

For determining the device of the spatial candidate motion vector being associated with the current portions of current video frame, wherein said Spatial candidate motion vector include the adjacent part neighbouring with described current portions for described current video frame and The motion vector determining；

For determining the device of the time candidate motion vector being associated with the described current portions of described current video frame, Wherein said time candidate motion vector includes the motion vector determining for the part of reference video frame；

For prune described spatial candidate motion vector with remove described spatial candidate motion vector in repetition person but The device of described time candidate motion vector is not considered during described pruning process；

For select described time candidate motion vector or prune after remaining described spatial candidate motion vector in One as selected candidate motion vector for the device of motion vector prediction process；

For remaining described spatial candidate motion vector after determining the described time candidate motion vector of identification and pruning List in described candidate motion vector in the device of index of position of described selected one；And

For sending the device of described index in bit stream with signal.

12. equipment according to claim 11, the wherein said device bag for pruning described spatial candidate motion vector Include for only pruning described spatial candidate motion vector to remove the described repetition in described spatial candidate motion vector The device of person.

13. equipment according to claim 11, its farther include for the spatial candidate motion determining described determination to Each in amount is through spatial prediction or the device through time prediction,

Wherein said for pruning described spatial candidate motion vector to remove the institute in described spatial candidate motion vector State the device of repetition person include for only prune described determination spatial candidate motion vector in be confirmed as through space Prediction spatial candidate motion vector, and do not prune described determination spatial candidate motion vector in be confirmed as through when Between prediction spatial candidate motion vector in the device of any one, and

Wherein said for select described time candidate motion vector or prune after remaining described spatial candidate motion One in Xiang Liang includes for selecting described time candidate to transport as the device of described selected candidate motion vector Moving vector, described spatial candidate motion vector in be confirmed as through time prediction one or pruning after remaining institute State in spatial candidate motion vector and be confirmed as the one through spatial prediction as described selected candidate motion vector Device.

14. equipment according to claim 11, it farther includes:

For determine described determination spatial candidate motion vector in each through spatial prediction still through time prediction Device；And

It is confirmed as through in the described spatial candidate motion vector of time prediction for replacing with default candidate motion vector The device of any one, wherein said default candidate motion vector comprises default motions vector information, wherein said silent Recognize motion vector information to include motion vector amplitude, identify that described reference frame was gone back in time before described present frame Prediction direction after being and the reference key identifying described reference frame, and

Wherein said for pruning described spatial candidate motion vector to remove the institute in described spatial candidate motion vector State the device of repetition person to include for pruning comprise in described default candidate motion vector one or more of Described spatial candidate motion vector is to remove the device of the described repetition person in described spatial candidate motion vector.

15. equipment according to claim 11, it farther includes that for determining be not through time prediction and to be different from institute In remaining described spatial candidate motion vector any one one after stating time candidate motion vector and pruning Or the device of more than one exceptional space candidate motion vector,

Wherein said for select described time candidate motion vector or prune after remaining described spatial candidate motion The device of the one in Xiang Liang include for select described time candidate motion vector, prune after remaining described sky Between one in candidate motion vector or the one in described exceptional space candidate motion vector as described selected The device of candidate motion vector.

16. 1 kinds of equipment for coding video frequency data, described equipment includes:

Motion compensation units, the spatial candidate motion vector that its determination is associated with the current portions of current video frame, its Described in spatial candidate motion vector include for described current video frame neighbouring with described current portions adjacent The motion vector partly determining；Determine the time candidate being associated with the described current portions of described current video frame Motion vector, wherein said time candidate motion vector include the motion that determines for the part of reference video frame to Amount；And prune described spatial candidate motion vector with remove described spatial candidate motion vector in repetition person but in institute Described time candidate motion vector is not considered during stating pruning process；

Mode selecting unit, after it selects described time candidate motion vector or prunes, remaining described spatial candidate is transported One in moving vector for motion vector prediction process and determines identification institute as selected candidate motion vector Described time in the list that after stating time candidate motion vector and pruning, remaining described spatial candidate motion is vectorial Select the index of the position of described selected one in motion vector；And

Entropy coding unit, it sends described index with signal in bit stream.

17. equipment according to claim 16, wherein said motion compensation units only prune described spatial candidate motion to Amount is to remove the described repetition person in described spatial candidate motion vector.

18. equipment according to claim 16,

Wherein said current portions includes current decoding unit CU,

The described part of wherein said reference frame includes the CU of described reference frame.

19. equipment according to claim 16, wherein said motion compensation units produces further and comprises prediction data Predicting unit PU, described prediction data includes at least described selected candidate motion vector,

Wherein said entropy coding unit sends described PU with signal in described bit stream.

20. equipment according to claim 16, wherein said entropy coding unit signal sends described index as motion Vector predictors MVP indexes, and described MVP index is surplus after identifying described time candidate motion vector and pruning The described position of the described selected candidate motion vector in the described list of remaining described spatial candidate motion vector.

21. equipment according to claim 16, wherein said entropy coding unit uses unitary code or truncates in unitary code One encodes described index to produce encoded index and to send described encoded rope with signal in described bit stream Draw.

22. equipment according to claim 16, wherein said mode selecting unit relative to described time Candidate Motion to After amount and pruning, each in remaining described spatial candidate motion vector performs rate-distortion analysis, and base Remaining described spatial candidate after described rate-distortion analysis selects described time candidate motion vector or prunes One in motion vector is as described selected candidate motion vector.

23. equipment according to claim 16,

Wherein said motion compensation units further determines that each in the spatial candidate motion vector of described determination is Through spatial prediction still through time prediction, and only prune the spatial candidate motion vector of described determination is confirmed as through The spatial candidate motion vector of spatial prediction, and do not prune in the spatial candidate motion vector of described determination and be confirmed as Through any one in the spatial candidate motion vector of time prediction, and

Wherein said mode selecting unit selects quilt in described time candidate motion vector, described spatial candidate motion vector Be defined as through time prediction one or pruning after remaining described spatial candidate motion vector in be confirmed as through The one of spatial prediction is as described selected candidate motion vector.

24. equipment according to claim 16, wherein said motion compensation units determines the spatial candidate fortune of described determination Each in moving vector through spatial prediction still through time prediction；Replace with default candidate motion vector and be confirmed as Through any one in the described spatial candidate motion vector of time prediction, wherein said default candidate motion vector comprises Default motions vector information, wherein said default motions vector information includes motion vector amplitude, identifies described reference The frame prediction direction before or after described present frame in time and the reference rope identifying described reference frame Draw；And prune comprise one or more of described spatial candidate motion in described default candidate motion vector to Amount is to remove the described repetition person in described spatial candidate motion vector.

25. equipment according to claim 16, wherein said motion compensation units further determines that it is not through time prediction And appointing in remaining described spatial candidate motion vector after being different from described time candidate motion vector and pruning One or more exceptional space candidate motion vectors of one,

After the described time candidate motion vector of wherein said mode selecting unit selection, pruning, remaining described space is waited Select the one in motion vector or the one in described exceptional space candidate motion vector as described selected candidate Motion vector.

26. 1 kinds of methods decoding video data, described method comprises:

Determining the spatial candidate motion vector being associated with the current portions of current video frame, wherein said spatial candidate is transported Moving vector includes for the spatially adjacent part neighbouring with the described current portions in described current video frame The nearby motion vectors determining；

Prune described spatial candidate motion vector with remove repetition person in described spatial candidate motion vector but described The time Candidate Motion not being considered for the described current portions of described current video frame during pruning process and determining Vector, wherein said time candidate motion vector includes the motion vector determining for the part of reference video frame；

Select described time Candidate Motion based in bit stream with the motion vector predictor candidates MVP index that signal sends After vector or pruning, the one in remaining described spatial candidate motion vector is as selected candidate motion vector For motion vector prediction process；And

Perform motion compensation based on described selected candidate motion vector.

27. methods according to claim 26, wherein prune described spatial candidate motion vector and include only pruning described sky Between candidate motion vector with remove described spatial candidate motion vector in described repetition person.

28. methods according to claim 26, it farther includes:

The number of candidate motion vector is defined as described time candidate motion vector plus described spatial candidate motion to The amount vector of remaining spatial candidate motion after trimming；

Dissect out decoded MVP index based on the number of the described determination of candidate motion vector from described bit stream, its Described in decoded MVP index include that the decoded MVP of unitary indexes and truncate unitary decoded MVP index In one；And

Decode described decoded MVP to index to determine that described MVP indexes.

29. methods according to claim 26, it farther includes:

Determine that the described time candidate motion vector of the described current portions of described present frame is unavailable；And

In response to determining that described time candidate motion vector is unavailable, determine the acquiescence fortune of described time candidate motion vector Moving vector information, wherein said default motions vector information include motion vector amplitude, identify described reference frame when Prediction direction before or after described present frame and the reference key identifying described reference frame between.

30. methods according to claim 29, it farther includes that the default motions vector information based on described determination is true The fixed context for performing the lossless statistical decoder of context-adaptive, wherein said context identification decoding table with With the described video data of decoding.

31. methods according to claim 29, wherein determine that described default motions vector information includes:

Determine described reference frame whether through intra-coding；And

When described reference frame is confirmed as through intra-coding, determine based on the described part for described reference frame Spatial motion vector obtains described default motions vector information.

32. methods according to claim 26, it farther includes:

Determine that the one in described spatial candidate motion vector is unavailable；And

Unavailable, based on motion vector prediction mode in response to the described one determining in described spatial candidate motion vector Determine and comprise to wait for the acquiescence of the default motions vector information of the described one in described spatial candidate motion vector Select motion vector, and

Wherein prune described spatial candidate motion vector to remove the described repetition person in described spatial candidate motion vector Including prune the one or more of described spatial candidate motion comprising in described default candidate motion vector to Measure to remove the described repetition person in described spatial candidate motion vector, and

Wherein select described time candidate motion vector, prune after remaining described spatial candidate motion vector in one Person as described selected candidate motion vector include based in described bit stream with signal send described motion to Amount predicted value MVP index selects in described time candidate motion vector or described spatial candidate motion vector by really One in remaining described spatial candidate motion vector after being set to disabled one or pruning.

33. methods according to claim 32, wherein when described motion vector prediction mode is that adaptive motion vector is pre- When surveying AMVP pattern, determine that described default motions vector information includes determining motion vector amplitude but uncertain knowledge The not described reference frame prediction direction before or after described present frame in time or identify described reference frame Reference key.

34. methods according to claim 32, wherein when described motion vector prediction mode is merging patterns, determine Described default motions vector information includes determining motion vector amplitude, identifies that described reference frame is worked as described in time Prediction direction before or after front frame and the reference key identifying described reference frame.

35. methods according to claim 32, it farther includes that the default motions vector information based on described determination is true The fixed context for performing the lossless statistical decoder of context-adaptive, wherein said context identification decoding table with With the described video data of decoding.

36. methods according to claim 26, it farther includes:

Determine that the one in described spatial candidate motion vector is unavailable；And

Unavailable in response to the described one determining in described spatial candidate motion vector, remove institute from described pruning process State in spatial candidate motion vector and be confirmed as disabled described one,

Wherein prune described spatial candidate motion vector to include only pruning in described spatial candidate motion vector and be confirmed as Available spatial candidate motion vector is with the described repetition person that removes in described spatial candidate motion vector but does not removes The time candidate motion vector determining for the described current portions of described current video frame or described spatial candidate Motion vector is confirmed as disabled described one.

37. methods according to claim 26, it further comprises determining that it is not through time prediction and when being different from described Between candidate motion vector and prune after remaining described spatial candidate motion vector in any one one or one Individual above exceptional space candidate motion vector,

After wherein selecting described time candidate motion vector or pruning in remaining described spatial candidate motion vector One includes selecting described time candidate motion vector, prune after in remaining described spatial candidate motion vector One in one or described exceptional space candidate motion vector is as described selected candidate motion vector.

38. 1 kinds of equipment being used for decoding video data, described equipment includes:

For determining the device of the spatial candidate motion vector being associated with the current portions of current video frame, wherein said Spatial candidate motion vector include the adjacent part neighbouring with described current portions for described current video frame and The motion vector determining；

For prune described spatial candidate motion vector with remove described spatial candidate motion vector in repetition person but The time candidate not being considered for the described current portions of described current video frame during described pruning process and determining The device of motion vector, wherein said time candidate motion vector includes determining for the part of reference video frame Motion vector；

For selecting described time candidate based in bit stream with the motion vector predictor candidates MVP index that signal sends After motion vector or pruning, the one in remaining described spatial candidate motion vector is as selected Candidate Motion Vector is for the device of motion vector prediction process；And

For performing the device of motion compensation based on described selected candidate motion vector.

39. equipment according to claim 38, the wherein said device bag for pruning described spatial candidate motion vector Include for only pruning described spatial candidate motion vector to remove the described repetition in described spatial candidate motion vector The device of person.

40. equipment according to claim 38, it farther includes:

Transport plus described spatial candidate for the number of candidate motion vector is defined as described time candidate motion vector The device of remaining spatial candidate motion vector after pruning in moving vector；

For dissecting out decoded MVP index based on the number of the described determination of candidate motion vector from described bit stream Device, wherein said decoded MVP index includes that the decoded MVP of unitary indexes and truncates unitary through translating One in code MVP index；And

Index for decoding described decoded MVP to determine the device that described MVP indexes.

41. equipment according to claim 38, it farther includes:

For determining the disabled device of described time candidate motion vector of the described current portions of described present frame；With And

In response to determining that described time candidate motion vector is unavailable, for determining the silent of described time candidate motion vector Recognizing the device of motion vector information, wherein said default motions vector information includes that motion vector amplitude, identification are described The reference frame prediction direction before or after described present frame in time and the ginseng identifying described reference frame Examine index.

42. equipment according to claim 41, it farther includes for the default motions vector letter based on described determination Breath determines the device of the context for performing the lossless statistical decoder of context-adaptive, wherein said context identification Decoding table is with in order to decode described video data.

43. equipment according to claim 41, the wherein said device bag for determining described default motions vector information Include:

For determining described reference frame whether through the device of intra-coding；And

It when described reference frame is confirmed as through intra-coding, is used for true based on the described part for described reference frame Fixed spatial motion vector obtains the device of described default motions vector information.

44. equipment according to claim 38, it farther includes:

For determining the disabled device of one in described spatial candidate motion vector；And

Unavailable, for based on motion vector prediction in response to the described one determining in described spatial candidate motion vector Pattern determination comprises the silent of the default motions vector information for the described one in described spatial candidate motion vector Recognize the device of candidate motion vector, and

Wherein said for pruning described spatial candidate motion vector to remove the institute in described spatial candidate motion vector State the device of repetition person to include for pruning comprise in described default candidate motion vector one or more of Described spatial candidate motion vector removing the device of described repetition person in described spatial candidate motion vector, and

Wherein said for select device include for select described time candidate motion vector, prune after remaining One in described spatial candidate motion vector as the device of described selected candidate motion vector include based on Described bit stream selects described time Candidate Motion with the described motion vector predictor candidates MVP index that signal sends Remaining described sky after being confirmed as disabled one or prune in vectorial or described spatial candidate motion vector Between the device of one in candidate motion vector.

45. equipment according to claim 44, wherein when described motion vector prediction mode is that adaptive motion vector is pre- When surveying AMVP pattern, the described device for determining described default motions vector information includes for determining motion Vector amplitude but the described reference frame of the uncertain identification prediction side before or after described present frame in time To or identify the device of reference key of described reference frame.

46. equipment according to claim 44, wherein when described motion vector prediction mode is merging patterns, described Device for determining described default motions vector information includes for determining motion vector amplitude, identifying described reference The frame prediction direction before or after described present frame in time and the reference rope identifying described reference frame The device drawing.

47. equipment according to claim 44, it farther includes for the default motions vector letter based on described determination Breath determines the device of the context for performing the lossless statistical decoder of context-adaptive, wherein said context identification Decoding table is with in order to decode described video data.

48. equipment according to claim 38, it farther includes:

For determining the disabled device of one in described spatial candidate motion vector；And

Unavailable in response to the described one determining in described spatial candidate motion vector, for moving from described pruning process It is confirmed as disabled described one in described spatial candidate motion vector,

The wherein said device for only pruning described spatial candidate motion vector includes waiting for only pruning described space Select the spatial candidate motion vector being confirmed as can use in motion vector to remove in described spatial candidate motion vector The time candidate fortune that determines for the described current portions of described current video frame of described repetition person but do not remove Moving vector or described spatial candidate motion vector are confirmed as the device of disabled described one.

49. equipment according to claim 38, it farther includes that for determining be not through time prediction and to be different from institute In remaining described spatial candidate motion vector any one one after stating time candidate motion vector and pruning Or the device of more than one exceptional space candidate motion vector,

Wherein said for select described time candidate motion vector or prune after remaining described spatial candidate motion The device of the one in Xiang Liang include for select described time candidate motion vector, prune after remaining described sky Between one in candidate motion vector or the one in described exceptional space candidate motion vector as described selected The device of candidate motion vector.

50. 1 kinds of equipment being used for decoding video data, described equipment includes:

Motion compensation units, the spatial candidate motion vector that its determination is associated with the current portions of current video frame, its Described in spatial candidate motion vector include determining for the adjacent part neighbouring with described current portions adjacent Motion vector；Prune described spatial candidate motion vector with remove repetition person in described spatial candidate motion but in institute The time candidate fortune not being considered for the described current portions of described current video frame during stating pruning process and determining Moving vector, wherein said time candidate motion vector include the motion that determines for the part of reference video frame to Amount；Described time candidate is selected to transport based in bit stream with the motion vector predictor candidates MVP index that signal sends Moving vector or prune after remaining described spatial candidate motion vector in one as selected Candidate Motion to Amount is for motion vector prediction process；And perform motion compensation based on described selected candidate motion vector.

51. equipment according to claim 50, wherein said motion compensation units only prune described spatial candidate motion to Amount is to remove the described repetition person in described spatial candidate motion vector.

52. equipment according to claim 50,

The number of candidate motion vector is defined as described time candidate motion vector and adds by wherein said motion compensation units Remaining spatial candidate motion vector after pruning in upper described spatial candidate motion vector, and

Wherein said equipment farther includes entropy decoding unit, and described entropy decoding unit is based on described in candidate motion vector The number determining dissects out decoded MVP index, wherein said decoded MVP index from described bit stream Including the decoded MVP of unitary indexes and truncates the one in unitary decoded MVP index；And decode described warp Decoding MVP indexes to determine that described MVP indexes.

53. equipment according to claim 50, it farther includes:

Determine that the described time candidate motion vector of the described current portions of described present frame is unavailable；And

In response to determining that described time candidate motion vector is unavailable, determine the acquiescence fortune of described time candidate motion vector Moving vector information, wherein said default motions vector information include motion vector amplitude, identify described reference frame when Prediction direction before or after described present frame and the reference key identifying described reference frame between.

54. equipment according to claim 53, it farther includes entropy decoding unit, and described entropy decoding unit is based on institute The default motions vector information stating determination determines for performing the context of the lossless statistical decoder of context-adaptive, its Described in context identification decoding table with in order to decode described video data.

55. equipment according to claim 53, whether wherein said motion compensation units further determines that described reference frame Through intra-coding, and when described reference frame is confirmed as through intra-coding, based on for locating in the reference frame In with described in the described reference frame of the described current portions position that location is identical in described present frame The spatial motion vector partly determining obtains described default motions vector information.

56. equipment according to claim 50, wherein said motion compensation units determines described spatial candidate motion vector In one unavailable；Unavailable, based on fortune in response to the described one determining in described spatial candidate motion vector Motion vector prediction pattern determine comprise for described spatial candidate motion vector in described one default motions to The default candidate motion vector of amount information；Prune comprise in described default candidate motion vector one or more Described spatial candidate motion vector with remove described spatial candidate motion vector in described repetition person；And based on Bit stream is selected in described time candidate motion vector by the motion vector predictor candidates MVP index that signal sends One, described spatial candidate motion vector in be confirmed as disabled one or prune after remaining described sky Between one in candidate motion vector.

57. equipment according to claim 56, wherein when described motion vector prediction mode is that adaptive motion vector is pre- When surveying AMVP pattern, described motion compensation units determines motion vector amplitude but the described reference frame of uncertain identification Prediction direction before or after described present frame or the reference key identifying described reference frame in time.

58. equipment according to claim 56, wherein when described motion vector prediction mode is merging patterns, described Motion compensation units determine motion vector amplitude, identify described reference frame in time before described present frame still Prediction direction afterwards and the reference key identifying described reference frame.

59. equipment according to claim 56, it farther includes entropy decoding unit, and described entropy decoding unit is based on institute The default motions vector information stating determination determines for performing the context of the lossless statistical decoder of context-adaptive, its Described in context identification decoding table with in order to decode described video data.

60. equipment according to claim 50, wherein said motion compensation units determines described spatial candidate motion vector In one unavailable；Unavailable in response to the described one determining in described spatial candidate motion vector, from described Pruning process removes in described spatial candidate motion vector and is confirmed as disabled described one；And only prune described The spatial candidate motion vector being confirmed as in spatial candidate motion vector can use is to remove described spatial candidate motion Described repetition person in Xiang Liang but do not remove the time determining for the described current portions of described current video frame Candidate motion vector or described spatial candidate motion vector are confirmed as disabled described one.

61. equipment according to claim 50, wherein said motion compensation units further determines that it is not through time prediction And appointing in remaining described spatial candidate motion vector after being different from described time candidate motion vector and pruning One or more exceptional space candidate motion vectors of one, and select described time candidate motion vector, repair In one during remaining described spatial candidate motion is vectorial after cutting or described exceptional space candidate motion vector One is as described selected candidate motion vector.